Software Engineer

throxy (yc p25); Greater London; On-site; GBP 90,000 - 130,000; Full time

Benefits offered by this job

Ownership

Senior peers

AI-native

Fast hiring

Impact from day one

Job summary

throxy is seeking a Software Engineer to join a fast-moving, AI-native team in London. You’ll own end-to-end delivery: design data models, migrations, production instrumentation, and metric monitoring.

The role requires strong TypeScript, web infra expertise, and experience shipping LLM features; you’ll work autonomously toward practical, shipped outcomes. In-person in London, Piccadilly, UK-based or relocating.

Qualifications

• 3+ years of production software engineering at a high-velocity company.
• Deep TypeScript and modern web infra. Includes Next.js, Bun, Postgres, ClickHouse, OpenSearch on AWS, with Datadog monitoring.
• Built systems that can handle real load in production.
• Shipped real LLM features: prompt design, evals, structured outputs, agentic workflows.
• Ability to work without PM handholding on ambiguous problems.
• Strong written communication; able to author design docs that align teams.
• Willingness to work in person in London, UK-based or relocate.

Responsibilities

• Design the data model, ship migrations, instrument production, and monitor metrics end-to-end.
• Own Maestro dialer and real-time calling infrastructure at scale.
• Develop the data engine: provider waterfalls, LLM eval pipelines, self-healing lead dataset, and scoring system.
• Build the intelligence layer: post-call insights and call-outcome classification.
• Replace external tools with in-house solutions where possible.

Skills

TypeScript

Web infra

End-to-end ownership

LLM features

Design docs

Bias to action

Written communication

Tools

Next.js

Bun

PostgreSQL

ClickHouse

OpenSearch

AWS

Datadog

Job description

About Throxy

throxy is a small applied research lab working on the future of sales.

Software Engineer @ throxy

throxy is a small applied research lab working on the future of sales.

We are a group of researchers, engineers, and technologists in London inventing at the edge of what's useful and possible. At throxy, we build agents that generate sales pipeline for our customers. Our customers don't care how it happens, they care that it does.

We're a small, talent-dense team moving extremely fast: 8 engineers shipping to production daily, talking directly to customers, and owning problems end-to-end.

The Role

You won't be handed tickets. You'll be handed a problem and trusted to scope it, sequence it, and ship it end-to-end: design the data model, ship the migrations, instrument production, watch the metric move.

What You'll Work On

• Maestro, our in-house dialer: real-time calling infrastructure on an Elixir core, placing millions of calls a month, with inbound, callbacks, and voicemail detection on local models
• The data engine underneath: provider waterfalls, LLM qualification and evaluation pipelines, a self-healing lead dataset, and scoring that decides who gets called next
• The intelligence layer: post-call insights and call-outcome classification
• Replacing every off-the-shelf tool we've outgrown with our own

How It Works Here

• Customers pay us for Meetings Booked, not for software. Being paid for the outcome deletes a whole class of problems: no engagement metrics, no A/B-tested button colors, no features promised by sales.
• No PMs. The roadmap comes from the funnel that feeds that KPI: pickup rate, time in conversation, data accuracy. You own a number, not a backlog
• Your users are an SDR floor making tens of thousands of dials a day on what you ship. Every dial is instrumented: a regression shows on a monitor in seconds, before anyone has to tell you. Their asks are triaged the same day, the urgent ones are in production before the evening

What We're Looking For

• 3+ years of production software engineering at a high-velocity company (we also hire at 6+ for senior scope)
• Deep TypeScript and modern web infra. We run Next.js, Bun, Postgres, ClickHouse, and OpenSearch on AWS, with Datadog over all of it
• You've built systems that had to hold up under real load
• You've shipped real LLM features: prompt design, evals, structured outputs, agentic workflows, and you know where they break
• Zero to one on ambiguous problems without a PM holding your hand
• Strong written communication: you can write a design doc that aligns a team
• Bias to action: you'd rather ship a working v1 today than a perfect v1 next month
• In person in London, 5 days a week (Piccadilly). UK-based or ready to relocate

Why throxy

• Real ownership: 8 engineers, big surface area, your fingerprints will be everywhere
• Senior bar: everyone you work with will sharpen you
• AI-native from day one, not retrofitting AI onto a legacy product
• Speed: our interview loop completes in 72 hours, and the job moves the same way
• Hit the ground running: you will have ownership of a hard problem within the first week